Introduction¶
Try to compare different algorithms on pure single cell results. Both the 10x and C1 SmartSeq sections are using the 81,881 entry ENCODE4 bulk annotation set.
Gene and Transcripts¶
Cell ranger and STAR Solo only produce Gene level count quantifications. When processing 10x data kallisto and salmon alevin produce gene counts as well.
When processing C1 data, Kallisto & Salmon only produce transcript quantifications, so for C1 gene level comparisons, transcript level quantifications are created by grouping all the transcripts by their gene id and summing those groups.
Correlations Used¶
This notebook reuses the correlation method developed for the ENCODE3 Bulk RNA pipeline
- Naive Spearman feeds all sample rows in raw data space to scipy's spearmanr function
- Rafa Spearman is a filtered spearman correlation that requires both values to be expressed (non-zero) and the average expression between the two samples to be > 0 in $log_2$. (So at least > 1 in the raw data space)
My implementation replicate_scores.
Original R version MAD.R
In [1]:
import numpy
import scanpy
import anndata
import os
import pandas
from matplotlib import pyplot
from matplotlib import cm
import scipy
from pathlib import Path
import upsetplot
import louvain
import sklearn
import seaborn
import logging
from pipeline_common import get_gene_id_to_name, load_gtf
#from cuml.decomposition import PCA
#from cuml.manifold import TSNE
In [2]:
logger = logging.getLogger('notebook')
In [3]:
analysis_dir = Path('ENCSR874BOF_e10_5_limb')
In [4]:
scanpy.__version__
Out[4]:
In [5]:
cellr_common = scanpy.read_h5ad(analysis_dir / 'cellranger_filtered.sparse.h5ad')
cellr_common
Out[5]:
In [6]:
solo_common = scanpy.read_h5ad(analysis_dir / 'solo_filtered.sparse.h5ad')
solo_common
Out[6]:
In [7]:
alevin_common = scanpy.read_h5ad(analysis_dir / 'alevin_filtered.h5ad')
alevin_common
Out[7]:
In [8]:
kallisto_common = scanpy.read_h5ad(analysis_dir / 'kallisto_filtered.h5ad')
kallisto_common
Out[8]:
In [9]:
def replicate_scores(rep1, rep2, rep1_name, rep2_name, Acutoff=0):
"""Compute correlations, MAD, and SD replicate comparison scores
Parameters
----------
table: pandas.DataFrame
DataFrame of RSEM quantifications
rep1_name: str
name of one column to score (library name)
rep2_name: str
name of the other column to score (library name)
Acutoff: float
Require that the average of the two replicates is greater
than this cutoff in log2 space
Returns
-------
pandas.Series
Containing all the scores between the two replicates
"""
#rep1 = table[rep1_name]
#rep2 = table[rep2_name]
eitherzero = (rep1 == 0) | (rep2 == 0)
replz1 = numpy.log2(rep1[eitherzero != True])
replz2 = numpy.log2(rep2[eitherzero != True])
M = replz1 - replz2
A = (replz1 + replz2) / 2.0
replz1_gt_Acutoff = replz1[A > Acutoff]
replz2_gt_Acutoff = replz2[A > Acutoff]
skip_rafa = False
if len(replz1_gt_Acutoff) < 100:
#logger.warning("No data survived Acutoff filter in %s", rep1_name)
skip_rafa = True
if len(replz2_gt_Acutoff) < 100:
#logger.warning("No data survived Acutoff filter in %s", rep2_name)
skip_rafa = True
if not skip_rafa:
rafa_pearson = scipy.stats.pearsonr(replz1_gt_Acutoff,
replz2_gt_Acutoff)[0]
rafa_spearman= scipy.stats.spearmanr(replz1_gt_Acutoff,
replz2_gt_Acutoff)[0]
else:
rafa_pearson = numpy.nan
rafa_spearman = numpy.nan
scores = pandas.Series({
'total_rows': len(rep1),
'passed_filter': len(replz1[A > Acutoff]),
'naive_pearson': scipy.stats.pearsonr(rep1, rep2)[0],
'naive_spearman': scipy.stats.spearmanr(rep1, rep2)[0],
'rafa_pearson': rafa_pearson,
'rafa_spearman': rafa_spearman,
'MAD': numpy.round(numpy.median(numpy.abs(M)[A > Acutoff]) * 1.4826, 3),
'SD': numpy.round(numpy.sqrt(numpy.mean(M[A > Acutoff] ** 2)), 3)
},
index=['total_rows', 'passed_filter',
'naive_pearson', 'naive_spearman',
'rafa_pearson', 'rafa_spearman',
'MAD', 'SD']
)
return scores
In [10]:
dense_10x = {
'cellranger': cellr_common.to_df().T,
'solo': solo_common.to_df().T,
'alevin': alevin_common.to_df().T,
'kallisto': kallisto_common.to_df().T,
#'kallisto EM': numpy.array(kallisto_em_common.X.todense()),
}
In [11]:
def compute_correlations(dense):
programs = list(dense.keys())
cell_correlations = {}
for name_x in programs:
for name_y in programs[programs.index(name_x):]:
assert dense[name_x].shape == dense[name_y].shape
for cell_id in dense[name_x].columns:
cs_cors = replicate_scores(dense[name_x][cell_id], dense[name_y][cell_id], name_x, name_y)
cell_correlations.setdefault(name_x, {}).setdefault(name_y, {})[cell_id] = cs_cors
return cell_correlations
Just a warning... the all vs all comparison is slow.
In [12]:
tenx_correlations = compute_correlations(dense_10x)
In [13]:
len(tenx_correlations['cellranger']['solo'])
Out[13]:
In [14]:
def plot_cell_correlation_histogram(table, metric, title=None, bins=50):
programs = list(table.keys())
cell_hists = {}
f = pyplot.figure(figsize=(7,7))
if title is not None:
f.suptitle(title.format(metric=metric))
plot_size = len(programs)-1
axes = f.subplots(plot_size, plot_size, sharex=True, sharey=True)
for x, name_x in enumerate(programs):
for y, name_y in enumerate(programs[programs.index(name_x)+1:]):
#plot_index = plot_size * (y+x) + x + 1
#ax = f.add_subplot(plot_size, plot_size, plot_index)
ax = axes[y+x, x]
if x == 0:
ax.set_ylabel(name_y)
spearman = []
for cell_id in table[name_x][name_y]:
spearman.append(table[name_x][name_y][cell_id][metric])
#spearman = numpy.asarray(spearman)
spearman = numpy.array(spearman)
spearman = spearman[~numpy.isnan(spearman)]
count = len(spearman)
median = numpy.median(spearman)
mean = numpy.mean(spearman)
cell_hists.setdefault(name_x, {})[name_y] = ax.hist(spearman, bins=bins, density=True)
ax.annotate(f'Mean {mean:0.2}\nMedian {median:0.2}\nCount {count}', xy=(0.1, 0.6), xycoords='axes fraction')
for y in range(plot_size):
axes[0, y].set_title(programs[y])
axes[plot_size-1, y].set_xlabel(programs[y])
#f.tight_layout()
In [15]:
def extract_score(correlations, metric, name_x, name_y):
score = {}
#name_x='cellranger'
#name_y='kallisto'
for cell_id in correlations[name_x][name_y]:
score[cell_id] = correlations[name_x][name_y][cell_id][metric]
return pandas.Series(score)
In [16]:
def sc_scatter(table, correlations, metric, name_x, name_y, cell_id, ax=None):
gridalpha = 0.5
def is_spike(x):
if x.startswith('gSpikein_') or x.startswith('tSpikein_'):
return True
else:
return False
if ax is None:
f = pyplot.figure()
ax = f.subplots(1,1)
set1 = cm.get_cmap('Set1').colors
colors = [{True: set1[0], False: set1[1]}[is_spike(x)] for x in table[name_x].index]
ax.plot([-5,10], [-5,10], c=set1[2])
ax.scatter(numpy.log2(table[name_x][cell_id]+0.01), numpy.log2(table[name_y][cell_id]+0.01), color=colors, s=2)
rafa_spearman = correlations[name_x][name_y][cell_id][metric]
passed_filter = correlations[name_x][name_y][cell_id]["passed_filter"]
ax.set_title(f'id {cell_id}\n{metric} {rafa_spearman:0.4}\nCount {int(passed_filter)}')
ax.set_xlabel(name_x)
ax.set_ylabel(name_y)
ax.grid(color='dimgrey', linestyle='-', linewidth=0.5, which="both", alpha = gridalpha)
In [17]:
def show_scatter_extremes(dense_mat, correlations, metric, name_x, name_y):
scores = extract_score(correlations, metric, name_x, name_y)
best_cell = scores.idxmax()
worst_cell = scores.idxmin()
# find smallest absolute difference from median and use that as median cell
median = numpy.abs(scores - scores.median())
median_cell = median.idxmin()
f = pyplot.figure(figsize=(12,4))
#f.suptitle(f'{metric} worst, median and best cells')
axes = f.subplots(1,3, sharex=True, sharey=True)
sc_scatter(dense_mat, correlations, metric, name_x, name_y, worst_cell, ax=axes[0])
sc_scatter(dense_mat, correlations, metric, name_x, name_y, median_cell, ax=axes[1])
sc_scatter(dense_mat, correlations, metric, name_x, name_y, best_cell, ax=axes[2])
Scatter plot comparisons¶
I have several slides where I wanted to get a sense of how the different algorithms performed on a cell.
What I settled on is a panel of 3 scatter plots that show the cell with the lowest correlation, the median correlation, and the best correlation between two selected algorithms.
Data points are in blue, spike-in values are in red and the line y=x is in green.
For 10x data there were no spike ins added, so all the spikes should be at 0,0 for naive, or min_threshold = $log_2 0.01$. Once in a while there's a read spuriously maps to the spike ins and so there might be one non-zero spike in the 10x data.
10x e10.5 full annotation¶
These plots are derived from ENCSR874BOF our e10.5 mouse forelimb dataset. They were analyzed with indicies built from the full ENCODE Bulk RNA-seq GTF file. This is unusual, many other 10x comparisons are done with a much smaller set that focuses on protein coding, lincRNA and antisense genes as well as the TR and IG annotations. A quick inspection suggests processed transcripts and pseudogenes appear to be the largest categories removed. See compare-10x-vs-ENCODE-gtf for more details.
The 10x cells selected were the intersection of Cellranger, STAR Solo, and Alevin methods, which happens to be the cells selected by STAR Solo.
10x full annotation naive spearman count correlation histogram (50 bins)¶
In [18]:
plot_cell_correlation_histogram(tenx_correlations, 'naive_spearman','Histogram of 10x per-cell {metric} correlations', bins=50)
10x full annotation worst, median, best naive spearman count cellranger vs star solo¶
In [19]:
show_scatter_extremes(dense_10x, tenx_correlations, 'naive_spearman', 'cellranger', 'solo')
10x full annotation correlations naive spearman worst, median, best cellrange vs alevin¶
In [20]:
show_scatter_extremes(dense_10x, tenx_correlations, 'naive_spearman', 'cellranger', 'alevin')
10x both naive spearman worst, median, best cellranger vs kallisto¶
In [21]:
show_scatter_extremes(dense_10x, tenx_correlations, 'naive_spearman', 'cellranger', 'kallisto')
10x naive spearman solo vs alevin¶
In [22]:
show_scatter_extremes(dense_10x, tenx_correlations, 'naive_spearman', 'solo', 'alevin')
10x spearman solo vs kallisto¶
In [23]:
show_scatter_extremes(dense_10x, tenx_correlations, 'naive_spearman', 'solo', 'kallisto')
10x spearman alevin vs kallisto¶
In [24]:
show_scatter_extremes(dense_10x, tenx_correlations, 'naive_spearman', 'alevin', 'kallisto')
10x rafa filtered count correlation histogram¶
In [25]:
plot_cell_correlation_histogram(tenx_correlations, 'rafa_spearman','Histogram of 10x per-cell {metric} correlations', bins=50)
10x counts worst, median, best rafa spearman cellranger vs solo¶
In [26]:
show_scatter_extremes(dense_10x, tenx_correlations, 'rafa_spearman', 'cellranger', 'solo')
In [27]:
show_scatter_extremes(dense_10x, tenx_correlations, 'rafa_spearman', 'cellranger', 'alevin')
In [28]:
show_scatter_extremes(dense_10x, tenx_correlations, 'rafa_spearman', 'cellranger', 'kallisto')
In [29]:
show_scatter_extremes(dense_10x, tenx_correlations, 'rafa_spearman', 'solo', 'alevin')
In [30]:
show_scatter_extremes(dense_10x, tenx_correlations, 'rafa_spearman', 'solo', 'kallisto')
In [31]:
show_scatter_extremes(dense_10x, tenx_correlations, 'rafa_spearman', 'alevin', 'kallisto')
C1 processed cells¶
In [32]:
suffix = '_gene_counts.h5ad'
c1_gene_counts = {}
for filename in sorted(Path('c1_pseudo/').glob('*'+suffix)):
name = filename.name[:-len(suffix)]
h5ad = scanpy.read_h5ad(filename)
c1_gene_counts[name] = h5ad.to_df().T
print(name, c1_gene_counts[name].shape)
In [33]:
c1_gene_count_correlations = compute_correlations(c1_gene_counts)
C1 naive spearman gene count histogram¶
In [34]:
plot_cell_correlation_histogram(c1_gene_count_correlations, 'naive_spearman','Histogram of per-cell {metric} C1 count correlations')
C1 gene counts worst, median, best both naive spearman kallisto vs rsem¶
In [35]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'kallisto', 'rsem')
C1 gene counts worst, median, best both non-zero kallisto vs salmon¶
In [36]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'kallisto', 'salmon_decoy')
C1 gene counts spearman kallisto vs salmon¶
In [37]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'kallisto', 'salmon')
C1 gene counts spearman kallisto vs star¶
In [38]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'kallisto', 'star')
C1 gene counts rsem vs salmon decoy¶
In [39]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'rsem', 'salmon_decoy')
C1 gene counts rsem vs salmon¶
In [40]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'rsem', 'salmon')
C1 gene counts spearman rsem vs star¶
In [41]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'rsem', 'star')
C1 gene counts spearman salmon_decoy vs salmon¶
In [42]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'salmon_decoy', 'salmon')
C1 gene counts salmon decoy vs star¶
In [43]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'salmon_decoy', 'star')
C1 gene counts salmon vs star¶
In [44]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'naive_spearman', 'salmon', 'star')
C1 gene counts rafa spearman histogram¶
In [45]:
plot_cell_correlation_histogram(c1_gene_count_correlations, 'rafa_spearman','Histogram of per-cell {metric} C1 count correlations')
In [46]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'kallisto', 'rsem')
In [47]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'kallisto', 'salmon_decoy')
In [48]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'kallisto', 'salmon')
In [49]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'kallisto', 'star')
In [50]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'rsem', 'salmon_decoy')
In [51]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'rsem', 'salmon')
In [52]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'rsem', 'star')
In [53]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'salmon_decoy', 'salmon')
In [54]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'salmon_decoy', 'star')
In [55]:
show_scatter_extremes(c1_gene_counts, c1_gene_count_correlations, 'rafa_spearman', 'salmon', 'star')
C1 Gene TPMs¶
In [56]:
suffix = '_gene_tpms.h5ad'
c1_gene_tpms = {}
for filename in sorted(Path('c1_pseudo/').glob('*'+suffix)):
name = filename.name[:-len(suffix)]
h5ad = scanpy.read_h5ad(filename)
c1_gene_tpms[name] = h5ad.to_df().T
print(name, c1_gene_tpms[name].shape)
In [57]:
c1_gene_tpm_correlations = compute_correlations(c1_gene_tpms)
Histogram C1 naive spearman gene TPMs¶
In [58]:
plot_cell_correlation_histogram(c1_gene_tpm_correlations, 'naive_spearman','Histogram of per-cell {metric} C1 gene TPM correlations', bins=50)
C1 gene TPMs best, median, worst both non-zero kallisto vs rsem¶
In [59]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'naive_spearman', 'kallisto', 'rsem')
In [60]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'naive_spearman', 'kallisto', 'salmon_decoy')
C1 gene TPMs best, median, worst naive spearman kallisto vs salmon¶
In [61]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'naive_spearman', 'kallisto', 'salmon')
C1 gene TPMs best, median, worst naive spearman rsem vs rsem¶
In [62]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'naive_spearman', 'rsem', 'salmon_decoy')
In [63]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'naive_spearman', 'rsem', 'salmon')
In [64]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'naive_spearman', 'salmon_decoy', 'salmon')
C1 gene TPM rafa spearman histogram¶
In [65]:
plot_cell_correlation_histogram(c1_gene_tpm_correlations, 'rafa_spearman','Histogram of per-cell {metric} C1 gene TPM correlations', bins=50)
In [66]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'rafa_spearman', 'kallisto', 'rsem')
In [67]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'rafa_spearman', 'kallisto', 'salmon_decoy')
In [68]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'rafa_spearman', 'kallisto', 'salmon')
In [69]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'rafa_spearman', 'rsem', 'salmon_decoy')
In [70]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'rafa_spearman', 'rsem', 'salmon')
In [71]:
show_scatter_extremes(c1_gene_tpms, c1_gene_tpm_correlations, 'rafa_spearman', 'salmon_decoy', 'salmon')
C1 transcript count naive spearman correlation histogram¶
In [72]:
suffix = '_transcript_counts.h5ad'
c1_transcript_counts = {}
for filename in sorted(Path('c1_pseudo/').glob('*'+suffix)):
name = filename.name[:-len(suffix)]
h5ad = scanpy.read_h5ad(filename)
c1_transcript_counts[name] = h5ad.to_df().T
print(name, c1_transcript_counts[name].shape)
In [73]:
c1_transcript_count_correlations = compute_correlations(c1_transcript_counts)
In [74]:
plot_cell_correlation_histogram(c1_transcript_count_correlations, 'naive_spearman','Histogram of per-cell {metric} C1 transcript count correlations', bins=50)
In [75]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'naive_spearman', 'kallisto', 'rsem')
In [76]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'naive_spearman', 'kallisto', 'salmon_decoy')
In [77]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'naive_spearman', 'kallisto', 'salmon')
In [78]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'naive_spearman', 'rsem', 'salmon_decoy')
In [79]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'naive_spearman', 'rsem', 'salmon')
In [80]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'naive_spearman', 'salmon_decoy', 'salmon')
C1 transcript count rafa spearman correlation histogram¶
In [81]:
plot_cell_correlation_histogram(c1_transcript_count_correlations, 'rafa_spearman','Histogram of per-cell {metric} C1 transcript count correlations', bins=50)
In [82]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'rafa_spearman', 'kallisto', 'rsem')
In [83]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'rafa_spearman', 'kallisto', 'salmon_decoy')
In [84]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'rafa_spearman', 'kallisto', 'salmon')
In [85]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'rafa_spearman', 'rsem', 'salmon_decoy')
In [86]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'rafa_spearman', 'rsem', 'salmon')
In [87]:
show_scatter_extremes(c1_transcript_counts, c1_transcript_count_correlations, 'rafa_spearman', 'salmon_decoy', 'salmon')
C1 transcript TPM naive spearman correlation histogram¶
In [88]:
suffix = '_transcript_tpms.h5ad'
c1_transcript_tpms = {}
for filename in sorted(Path('c1_pseudo/').glob('*'+suffix)):
name = filename.name[:-len(suffix)]
h5ad = scanpy.read_h5ad(filename)
c1_transcript_tpms[name] = h5ad.to_df().T
print(name, c1_transcript_tpms[name].shape)
In [89]:
c1_transcript_tpm_correlations = compute_correlations(c1_transcript_tpms)
In [90]:
plot_cell_correlation_histogram(c1_transcript_tpm_correlations, 'naive_spearman','Histogram of per-cell {metric} C1 transcript TPM correlations', bins=50)
In [91]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'naive_spearman', 'kallisto', 'rsem')
In [92]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'naive_spearman', 'kallisto', 'salmon_decoy')
In [93]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'naive_spearman', 'kallisto', 'salmon')
In [94]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'naive_spearman', 'rsem', 'salmon_decoy')
In [95]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'naive_spearman', 'rsem', 'salmon')
In [96]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'naive_spearman', 'salmon_decoy', 'salmon')
C1 transcript tpm rafa spearman correlation histogram¶
In [97]:
plot_cell_correlation_histogram(c1_transcript_tpm_correlations, 'rafa_spearman','Histogram of per-cell {metric} C1 transcript TPM correlations', bins=50)
In [98]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'rafa_spearman', 'kallisto', 'rsem')
In [99]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'rafa_spearman', 'kallisto', 'salmon_decoy')
In [100]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'rafa_spearman', 'kallisto', 'salmon')
In [101]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'rafa_spearman', 'rsem', 'salmon_decoy')
In [102]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'rafa_spearman', 'rsem', 'salmon')
In [103]:
show_scatter_extremes(c1_transcript_tpms, c1_transcript_tpm_correlations, 'rafa_spearman', 'salmon_decoy', 'salmon')
